Misc LargestContentfulPaint updates

[tunetheweb]

Summary

Correct erroneous LargestContentfulPaint data I spotted:

• Safari and Firefox do not support presentationTime
• Safari DOES support cross-origin render times.
• Chrome does support paintTime and presentationTime but behind a flag (unflagging in Chrome 145!)

Test results and supporting details

Safari does support cross-origin lcp render time but not presentationTime (tested on https://www.tunetheweb.com/experiments/lcp-cross-origin/)

Firefox gives null for presentationTime:

Chrome 134 supports paintTime and presentationTime when experimental web platform features are turned on:

Related issues

[github-actions]

Tip: Review these changes grouped by change (recommended for most PRs), or grouped by feature (for large PRs).

[Elchi3]

The collector detected support and it seems it is exposed. The null behavior seems to be per spec also? Maybe we should mark it as partial? (this is always weird for feature detection but oh well).

Suggested change

	"version_added": false
	"version_added": "140",
	"partial_implementation": true,
	"notes": "This property is exposed, but always returns `null`."

[tunetheweb]

It is effectively allowed per spec not to support this I agree, so it is an optional feature. But this is where it's even more important for MDN to document this IMHO.

Also I agree the fact it's there (but always null) could be interpreted as having "partial support". But to me, since it's always null, it's equivalent to "no support" rather than partial.

WDYT @canova from an Firefox side? And @smfr from a Safari size? Do you think it's fair to say you have not implemented presentationTime at this time for LCP? Note Interop and Baseline definitions do not include this key, and I'm keen to keep that this way, since these are optional, so you would still be marked as supporting LCP in those.

[tunetheweb]

Actually, let's go with partial support for now, to unblock this, since there are other updates in this PR I'd like to have sooner rather than later and partial is certainly better than the current reading of them having full support. But curious to hear @canova and @smfr's opinion and can follow up with another PR if necessary.

[Elchi3]

Same here?

[tunetheweb]

Suggested change

	"version_added": false
	"version_added": "26.2",
	"partial_implementation": true,
	"notes": "This property is exposed, but always returns `null`."

[Elchi3]

Thanks @tunetheweb, let's go with partial for now.